- Email: [email protected]
Acoustic analysis of voice in multiple sclerosis patients
Acoustic Analysis of Voice in Multiple Sclerosis Patients *Adriana Ve´lez Feijo´, *†Maria Alice Parente, ‡Mara Behlau, §||Se´rgio Haussen, †Maria Cecı´lia De Veccino, and †Beatriz Castellar de Faria Martignago Porto Alegre, RS, Brazil
Summary: The objective is to investigate the presence of dysphonic symptoms in multiple sclerosis (MS) patients and to compare quantitative acoustic parameters in multiple sclerosis patients and normal individuals. The method of study was an 8-month controlled cross-sectional that was carried out with 106 individuals (30 MS, 76 controls). Both groups included males and females from 20 to 55 years. Exclusion criteria were prior vocal disorder, laryngeal microsurgery, recent endotracheal intubation, tumors, laryngeal, lung or mediastinal metastases, respiratory disease, and other associated neurological diagnoses. For dysphonic symptoms (qualitative variables), associations were assessed using Mantel-Haenszel’s χ2test, with Yates correction or the Fisher exact test when necessary. Statistical significance was set at p ⭐ 0.05. Dysphonia was observed in 70% of MS individuals versus 33% of controls (p ⫽ 0.01). Association was found between MS and dysphonia (OR: 2.2, CI 95%: 1.13–4.25). Fundamental frequency was higher among MS patients (p ⫽ 0.01). Fundamental frequency deviation was significantly higher in MS women (but not men) than controls (p ⫽ 0.00). Jitter was higher in MS men than in all other groups (p ⫽ 0.00). Results suggest that evaluation and treatment of MS patients should be revised, evaluating voice alterations in relation to other signs. MS seems to intensify gender effect on fundamental frequency deviation, noise, and jitter, with MS women presenting fewer voice variations than men. Key Words: Voice—Multiple sclerosis—Hoarseness—Speech acoustics.
INTRODUCTION
Accepted for publication May 6, 2003. From the *Universidade Federal do Rio Grande do Sul; †Psychology Department, Psychology Institute, Universidade Federal do Rio Grande do Sul; ‡Centro de Estudos da Voz (CEV), Sa˜o Paulo; §Fundac¸a˜o Faculdade Federal de Cieˆncias Me´dicas de Porto Alegre (FFFCMPA); ||Neurology Service, Irmandade da Santa Casa de Miserico´rdia. Address correspondence and reprint requests to Adriana Ve´lez Feijo´, Universidade Federal do Rio Grande do Sul, Av Nilo Pec¸anha 550 apto 804, 90470-000 Porto Alegre, RS, Brazil. E-mail: [email protected] Journal of Voice, Vol. 18, No. 3, pp. 341–347 0892-1997/$30.00 쑕 2004 The Voice Foundation doi:10.1016/j.jvoice.2003.05.004
Multiple sclerosis (MS), initially described by Charcot, in 1877,1 is a chronic degenerative disease that affects the myelin sheath of the central nervous system (CNS) and is characterized by multiple lesions of brain white matter, brainstem, and spinal cord, so that several motor-functional systems may be involved, such as the pyramidal, cerebellar, and brainstem systems. Scanning speech has been considered as one of the cardinal symptoms of MS.2,3 Other clinical signs include spasticity, strength deficit, ataxia, language disorders, cognitive and emotional alterations, tremor, and disorders of the 341
´ ET AL ADRIANA VE´LEZ FEIJO
342
autonomic nervous system and of the cranial nerves, such as changes in vision and ocular motility, dysphagia, and dysarthria.4–10 Dysarthria is a motor speech disorder that may be associated with alterations in the regulation of phonation. The phonatory symptoms in MS are often masked by the pronounced scanning speech and abnormal prosodic patterns while speaking. However, the voice is a highly sensitive neurophysiological function, and a detailed study of vocal measures may reflect the status of nervous centers and pathways.7 Therefore, the study of the voice in carriers of neurological disorders aims at identifying relationships among pathophysiological mechanisms, lesion topography, and distinct phonatory alterations, for example, dysphonic symptoms. In MS, this type of finding may contribute toward maximizing the functional abilities of affected individuals. The objective of the present study was (1) to assess the presence of dysphonic symptoms in multiple sclerosis patients and (2) to compare quantitative acoustic parameters (fundamental frequency, jitter, shimmer, standard deviation of fundamental frequency, glottal noise level, frequency, and amplitude of tremor) in multiple sclerosis patients and normal individuals.
Only MS patients in the relapsing-remitting category as defined by Poser et al11 were included in the study. Exclusion criteria included the presence of vocal disorder prior to the appearance of neurological or clinical symptoms, previous larynx microsurgery, recent episode of endotracheal intubation, primary or metastatic tumor of the larynx, lung, or mediastinum, diagnosis of respiratory disease (acute/chronic), and any associated neurological diagnosis. The difference in gender distribution between the groups was statistically controlled. The study protocol was approved by the Ethics Committee at HCPA. All patients signed a consent form after receiving information about the objectives of the study and about the procedures to be carried out. During the data collection period, monthly meetings were held with the medical team in charge of the MS patients to review diagnostic criteria and to clarify possible doubts regarding inclusion of new patients. The acoustic evaluation of each patient was individually reviewed after the sample was recorded. Control of environmental noise was carried out following standard procedures.12 Data were entered into the database twice; a second analysis was carried out for 20% of the subjects, selected at random.
METHODS
Measurement of study variables A vocal sample was collected using a microphone (Optimus unidirectional microphone IMP.600, Philippines, 33-3017) connected to a computer (Acer, Extensa 368) through a preamplifier (Tiger Electronics, Seattle, WA). A computerized acoustic analysis of the samples was carried out using the software Dr. Speech Sciences, Version 3.2 (Tiger Electronics, Seattle, WA). This software was used to obtain quantitative parameters and graphic representations. For collection of vocal samples, individuals were asked to emit a sustained vowel /a / (as in father), trying to maintain the usual frequency and intensity. The microphone was placed on a microphone stand at a distance of 10 cm from the individual’s mouth. The rate used to record the signal was 44,100 Hz. In the present study, dysphonic symptoms were assessed using spectrographic analysis of wide and narrow bands. In the wideband analysis, the following items were scored: peak periodicity, darkness of
An 8-month controlled, cross-sectional study was designed to assess dysphonic symptoms in MS. The sample was composed of 30 individuals with multiple sclerosis, selected from among the patients in follow-up at the neurology service at Hospital de Clı´nicas de Porto Alegre (HCPA), a teaching hospital in southern Brazil. A control group was set up with 76 HCPA employees without neurological manifestations or psychiatric diagnoses, and without any chronic organic diseases. Socioeconomic status was similar for the MS and the control groups. Age ranged from 20 to 55 years. To calculate the size of the sample, the following criteria were observed: significance level (α) of 5%; power of 80% (β); and a control-case ratio of 2:1. This ratio was established because we expected dysphonic symptoms to be more frequent in the MS population in relation to controls. The exposure of normal individuals was estimated at 15%. Journal of Voice, Vol. 18, No. 3, 2004
MULTIPLE SCLEROSIS PATIENTS the contour, and presence of noise in high frequencies and in the formant region. In the narrow band, the presence of harmonics up to 4000 Hz as well as the regularity and definition of the contour were assessed. If three out of four items were altered in the wideband analysis, the examination was classified as pathologic; in the narrowband analysis, two out of three were altered. The spectrograms for the two bands were later compared. Dysphonic symptoms were diagnosed when both examinations were classified as pathologic.13 Assessment of the examinations was performed by a voice specialist. The examiner was blinded for the presence or absence of the study factor (MS). After that, fundamental frequency, jitter, shimmer, and noise measures were extracted from the samples. The software employed in the present study (Dr. Speech Sciences, Version 3.0)12 defines as normal jitter values up to 0.5% and shimmer values up to 3.0%. Glottal noise is considered to be present at an acoustic spectral level of ⫺10 dB. A maximum standard deviation of 3.0 Hz in relation to frequency is considered. The reference values for frequency were 80 to 150 Hz for males and 150 to 250 Hz for females.14 Statistical analysis The statistical analysis was carried out using the Statistical Package for the Social Sciences (SPSS, Chicago, IL).15 For dysphonic symptoms (qualitative variables), associations were assessed using Mantel-Haenszel’s χ2test, with Yates correction or the Fisher exact test when necessary. A p ⭐ 0.05 was considered to be significant. Jitter, shimmer, glottal noise, fundamental frequency, and frequency deviation (continuous variables) were analyzed using ANOVA or Student t test for independent samples or for dependent samples, when necessary. Multivariate analysis (MANOVA) was used to control confounding variables. Finally, a logistic regression model was used to analyze dependent variables both for MS and the presence of dysphonic symptoms. The model included the independent variables that presented differences in the MANOVA or χ2univariate analysis. RESULTS The spectrographic analysis of wide and narrow bands revealed that the frequency of dysphonic
343
symptoms among MS patients was 70% (n ⫽ 21/30) and, among controls, 33% (n ⫽ 25/76) (p ⫽ 0.01). Figure 1 shows the wideband and narrowband spectrogram of an MS patient with dysphonic symptoms in comparison with a normal individual. The demographic characteristics of the sample and the results of acoustic analysis appear in Table 1. Fundamental frequency and fundamental frequency deviation were higher in MS patients in comparison with the control group (p ⫽ 0.01). Because of the known importance of the gender variable for several aspects of the voice, the impact of this variable was analyzed separately. The controlled analysis of gender in individuals with and without multiple sclerosis is presented in Table 2. Jitter values were higher in MS men than in all other groups. Fundamental frequency deviation was higher in normal men than in normal women. In addition, this parameter was almost twice as high in MS men when compared with MS women, who presented values that were similar to those of normal women. The same was observed for glottal noise. To enable a more dynamic assessment of the simultaneous effects of the variables under study, a logistic regression model was set up having exposure to MS as the dependent variable, and gender, dysphonic symptoms, fundamental frequency, shimmer, noise, and fundamental frequency deviation as independent variables. After this analysis, dysphonic symptoms and fundamental frequency remained in the model (Table 3). From these, dysphonic symptoms presented a significant association with the dependent variable exposure to MS. Thus, the odds for having MS were 2.2 times higher in individuals presenting dysphonic symptoms.
DISCUSSION In the present study, we observed a statistically significant association between MS and dysphonic symptoms, with a frequency of 70% among subjects found to have MS. This finding is most probably explained by alterations in periaqueductal gray matter, which are common in MS.3,5 Such alterations have a major impact on phonation and are characterized by neuronal concentration and projections of abdominal, intercostal, pharyngeal, palatal, lingual, Journal of Voice, Vol. 18, No. 3, 2004
344
´ ET AL ADRIANA VE´LEZ FEIJO
FIGURE 1. Left. Narrowband analysis. Narrowband spectrogram of an MS patient with dysphonic symptoms (A) in comparison with a normal individual (B). Observe the lack and irregularity of harmonics in the MS patient. Right. Wideband analysis. Wideband spectrogram of an MS patient with dysphonic symptoms (C) in comparison with a normal individual (D). Observe the aperiodicity and a lack of darkness in the contour of the MS patient.
facial, masticatory, and laryngeal motoneurons into the nucleus retroambigualis.14,16,17 Even though the jitter values for both MS and normal men were within the normal range, our results suggest that the effect of jitter may be increased in MS men. The variability of muscle contraction resulting from the triggering of the motor unit in MS is one possible explanation for this. It is important to stress that chemical and electrical discharges, as well as the process of recovery of the myoneural junction, are faster than the recovery of the mechanical process.18 Therefore, the slowing down of the electrical impulse through the demyelination process Journal of Voice, Vol. 18, No. 3, 2004
in MS may cause an alteration of the recovery synchrony between the aspects mentioned previously, leading to a larger variation in the cycle-to-cycle frequency, ie, short-term variation or jitter. However, this finding must be interpreted with care, because jitter was not different when women with and without MS were compared. Also, it is important to stress that jitter, shimmer, and noise were not associated with MS in the logistic regression model, as also reported in other studies.7,19 Regarding long-term phonatory instability in MS subjects, both tremor amplitude and frequency were within the normal range for both genders. These
MULTIPLE SCLEROSIS PATIENTS TABLE 1. Demographic Information and Results of Vocal Analysis§ in Patients With and Without Multiple Sclerosis
Variables Age Gender Male Female Frequency Jitter Shimmer Fundamental frequency deviation Glottal noise Tremor frequency Tremor amplitude
Multiple sclerosis n ⫽ 30
Controls n ⫽ 76
P
45.10 ⫾ 7.24
42.16 ⫾ 7.64
0.07*
9.00 21.00 175.53 ⫾ 39.68 0.26 ⫾ 0.17 2.94 ⫾ 1.84 2.30 ⫾ 0.95
53.00 23.00 150.71 ⫾ 39.05 0.24 ⫾ 0.09 2.58 ⫾ 1.06 1.61 ⫾ 0.87
0.01** 0.01* 0.27* 0.21* 0.01*
⫺12.37 ⫾ 4.45 1.72 ⫾ 0.94
⫺12.78 ⫾ 3.56 2.96 ⫾ 3.87
0.62* 0.09*
1.44 ⫾ 0.69
1.47 ⫾ 1.03
0.90*
*
One-way variance analysis or MANOVA. χ test with Yates correction. § Sound card capture. ** 2
results differ from those of previous studies reporting tremor values starting at 11.2 dB and 7.5 Hz.20 However, the present results may have been affected by limitations in the software used for assessment, because in general these measurements did not match our clinical observations. For example, tremor
345
frequency and amplitude values were altered even in individuals with a normal voice and with a normal spectrographic evaluation. In the present study, all MS patients belonged to the relapsing-remitting category11,21; therefore, the assessment of voice was not related to degree of disability. However, it is possible to observe that the mean time since diagnosis was similar for men and women with MS. Despite the fact that MS patients were not individually assessed in terms of degree of disability, the highest possible number of individuals was enrolled during the 8 consecutive months of this study; the subjects enrolled presented a comparable disease duration, which characterized a similar profile. In the spectrographic analysis, certain contour characteristics were observed, such as irregular harmonics, presence of noise, and reduced darkening of the contour. Previous studies report that dysphonia is a symptom rather than a disease per se.6,7,22 In the case of neurological disorders, dysphonic symptoms are often among the earliest symptoms developed. We believe that the strong association between dysphonic symptoms and MS described in the present study may help to demonstrate that health professionals must be alert to the need for patients to be referred for neurological examination, as well as changing the cultural and social values that define a normal voice. In addition, our results underscore the
TABLE 2. Analysis of the Interaction Between Variables According to Gender and Presence or Absence of Multiple Sclerosis* Multiple sclerosis (n ⫽ 30)
Controls (n ⫽ 76)
Variable
Males (n ⫽ 9)
Females (n ⫽ 21)
Males (n ⫽ 53)
Females (n ⫽ 23)
F
P
Age Disease duration Fundamental frequency Jitter Shimmer Fundamental frequency deviation Glottal noise Tremor frequency Tremor amplitude
47.11 ⫾ 3.52 6.63 ⫾ 4.87 142.43 ⫾ 27.36
44.24 ⫾ 8.28 7.40 ⫾ 6.69 189.71 ⫾ 35.78
41.40 ⫾ 7.83 – 133.58 ⫾ 27.34
43.91 ⫾ 7.02 – 191.96 ⫾ 31.64
29.5399
0.00
0.33 ⫾ 0.30 2.02 ⫾ 1.08 4.11 ⫾ 2.79
0.23 ⫾ 0.05 2.43 ⫾ 0.89 2.44 ⫾ 0.96
0.23 ⫾ 0.08 1.36 ⫾ 0.65 2.69 ⫾ 1.17
0.25 ⫾ 0.10 2.20 ⫾ 1.04 2.33 ⫾ 0.67
4.6453 2.2033 10.6975
0.00 0.09 0.00
⫺9.89 ⫾ 4.51 2.02 ⫾ 1.07 1.51 ⫾ 0.87
⫺13.44 ⫾ 0.08 1.59 ⫾ 0.88 1.41 ⫾ 0.62
⫺11.92 ⫾ 3.72 2.58 ⫾ 3.25 1.54 ⫾ 1.20
⫺14.84 ⫾ 1.99 3.87 ⫾ 5.03 1.28 ⫾ 0.38
5.5792 1.8272 0.4161
0.00 0.14 0.74
Significance level ⬍0.05. Repeated letters in the same parameter indicate a significant difference between the groups.
*
Journal of Voice, Vol. 18, No. 3, 2004
´ ET AL ADRIANA VE´LEZ FEIJO
346
TABLE 3. Results of the Logistic Regression Analysis for Exposure to Multiple Sclerosis* b
Standard error
⫺0.7861 ⫺0.0164 4.0501
0.3369 0.0085 1.5329
Variables not in model
Score
p
Gender Shimmer Noise Fundamental frequency deviation
0.9307 0.5559 0.9959 0.0609
Variables in model Dysphonic symptoms Frequency Constant
0.33 0.45 0.31 0.80
p 0.01 0.05 0.00
R
OR
CI 95%
⫺0.1656 ⫺0.1176
2.2 1.0
1.13–4.25 0.97–1.00
R 0.0000 0.0000 0.0000 0.0000
Partial correlation is expressed as R and ranges from ⫺1 to ⫹1. OR ⫽ odds ratio. CI 95% ⫽ 95% confidence interval.
*
consequences that disregarding small vocal alterations may have for the health of affected individuals. This finding may also contribute toward speeding the process of diagnosis of neurological diseases.
CONCLUSIONS Our results reveal an association between MS and dysphonic symptoms. MS tends to worsen performance of the acoustic parameters fundamental frequency deviation, noise, and jitter. Men with MS present with worse vocal performance than do women with MS. The phoniatric therapeutic process in MS patients should be reevaluated in terms of focusing on the phonation and resonance levels of the phonatory system, and in terms of adding vocal therapy methods, because of the large association observed between dysphonic symptoms and MS. This would enable more favorable communication conditions in these patients, and it would mitigate the limitations associated with MS.
REFERENCES 1. Charcot J. Lectures on the diseases of the nervous system. London: The New Sudenhem Society; 1877:1. 2. Bauer HJ, Hanefeld FA. Multiple sclerosis—Its impact from childhood to old age. London: Saunders; 1993:132–170. 3. Matthews WB, Compston A, Allen IV, Martyn CN. McAlpine’s multiple sclerosis. 2nd ed. London: Churchill; 1991: 43–105. Journal of Voice, Vol. 18, No. 3, 2004
4. Darley F, Brown J, Goldstein N. Dysarthria in multiple sclerosis. J Speech Hear Res. 1972;15:229–245. 5. Guberman A. An introduction to clinical neurology: Pathophysiology, diagnosis, and treatment. Boston: Little, Brown; 1994. 6. Hartelius L, Svensson P. Speech and swallowing symptoms associated with Parkinson’s disease and multiple sclerosis: a survey. Folia Phoniatr Logop. 1994;46:9–17. 7. Ramig AL, Titze I, Scherer R, Ringel S. Acoustic analysis of voice of patients with neurologic disease: rationale and preliminary data. Ann Otol Rhinol Laryngol. 1988;97: 164–171. 8. Poser CM. The epidemiology of multiple sclerosis: a general overview. Ann Neurol. 1994;36:180–193. 9. Beukelman DR, Kraft GH, Freal J. Expressive communication disorders in persons with multiple sclerosis: a survey. Arch Phys Med Rehabil. 1985;10:675–677. 10. Lethlean J, Murdoch B. Naming in multiple sclerosis: effects of disease course, duration, age, and education level. J Med Speech Lang Pathol. 1994;2(1):43–56. 11. Poser CM, Paty DM, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol. 1983;13:227–231. 12. Huang D, Minifie FD. Dr. Speech science for Windows software and real speech—user’s guide. Seattle, WA: Tiger Electronics US Inc.; 1995. 13. Rontal E, Rontal M, Rolnick MI. Objective evaluation of vocal pathology using voice spectrography. In: Baken RJ, Daniloff RG, eds. Readings in clinical spectrography of speech. California: Singular; 1991554–563. 14. Norusis MJ. Statistical package for the social sciences SPSS (SPSS/PC) for IBM PC/XT/AT. Chicago, IL: SPSS Inc.; 1986. 15. Hammarberg B, Gauffin J. Perceptual and acoustic characteristics of quality differences in pathological voices as related to physiological aspects. In: Fujimura O, Hirano M, eds. Vocal fold physiology. California: Singular; 1995: 238–303.
MULTIPLE SCLEROSIS PATIENTS 16. Holstege G, Ehling T. Two motor systems involved in the production of speech. In: Davis PJ, Fletcher NH, eds. Vocal fold physiology—controlling complexity and chaos. California: Singular; 1996:153–169. 17. Zhang SP, Davis PJ, Bandler R. Brain stem integration of vocalization: role of the nucleus retroambigualis. J Neurophysiol. 1995;74:2500–2512. 18. Ludlow C, Bassich C, Connor N, Coulter D. Phonatory characteristics of vocal fold tremor. J Phonet. 1986;14: 509–515.
347
19. Titze IR. A model for neurologic sources of aperiodicity in vocal fold vibration. J Speech Hear Res. 1991;34:460–472. 20. Davis P, Zhang SP, Bandler R. Midbrain and medullary regulation of vocalization. In: Davis PJ, Fletcher NH, eds. Vocal fold physiology—controlling complexity and chaos. California: Singular; 1996:121–134. 21. Hanson D. Neuromuscular disorders of the larynx. Otolaryngol Clin North Am. 1991;24:1035–1051. 22. Zwirner P, Murry T, Woodson G. Phonatory function of neurologically impaired patients. J Commun Disord. 1991; 24:287–300.
Journal of Voice, Vol. 18, No. 3, 2004
Summary: The objective is to investigate the presence of dysphonic symptoms in multiple sclerosis (MS) patients and to compare quantitative acoustic parameters in multiple sclerosis patients and normal individuals. The method of study was an 8-month controlled cross-sectional that was carried out with 106 individuals (30 MS, 76 controls). Both groups included males and females from 20 to 55 years. Exclusion criteria were prior vocal disorder, laryngeal microsurgery, recent endotracheal intubation, tumors, laryngeal, lung or mediastinal metastases, respiratory disease, and other associated neurological diagnoses. For dysphonic symptoms (qualitative variables), associations were assessed using Mantel-Haenszel’s χ2test, with Yates correction or the Fisher exact test when necessary. Statistical significance was set at p ⭐ 0.05. Dysphonia was observed in 70% of MS individuals versus 33% of controls (p ⫽ 0.01). Association was found between MS and dysphonia (OR: 2.2, CI 95%: 1.13–4.25). Fundamental frequency was higher among MS patients (p ⫽ 0.01). Fundamental frequency deviation was significantly higher in MS women (but not men) than controls (p ⫽ 0.00). Jitter was higher in MS men than in all other groups (p ⫽ 0.00). Results suggest that evaluation and treatment of MS patients should be revised, evaluating voice alterations in relation to other signs. MS seems to intensify gender effect on fundamental frequency deviation, noise, and jitter, with MS women presenting fewer voice variations than men. Key Words: Voice—Multiple sclerosis—Hoarseness—Speech acoustics.
INTRODUCTION
Accepted for publication May 6, 2003. From the *Universidade Federal do Rio Grande do Sul; †Psychology Department, Psychology Institute, Universidade Federal do Rio Grande do Sul; ‡Centro de Estudos da Voz (CEV), Sa˜o Paulo; §Fundac¸a˜o Faculdade Federal de Cieˆncias Me´dicas de Porto Alegre (FFFCMPA); ||Neurology Service, Irmandade da Santa Casa de Miserico´rdia. Address correspondence and reprint requests to Adriana Ve´lez Feijo´, Universidade Federal do Rio Grande do Sul, Av Nilo Pec¸anha 550 apto 804, 90470-000 Porto Alegre, RS, Brazil. E-mail: [email protected] Journal of Voice, Vol. 18, No. 3, pp. 341–347 0892-1997/$30.00 쑕 2004 The Voice Foundation doi:10.1016/j.jvoice.2003.05.004
Multiple sclerosis (MS), initially described by Charcot, in 1877,1 is a chronic degenerative disease that affects the myelin sheath of the central nervous system (CNS) and is characterized by multiple lesions of brain white matter, brainstem, and spinal cord, so that several motor-functional systems may be involved, such as the pyramidal, cerebellar, and brainstem systems. Scanning speech has been considered as one of the cardinal symptoms of MS.2,3 Other clinical signs include spasticity, strength deficit, ataxia, language disorders, cognitive and emotional alterations, tremor, and disorders of the 341
´ ET AL ADRIANA VE´LEZ FEIJO
342
autonomic nervous system and of the cranial nerves, such as changes in vision and ocular motility, dysphagia, and dysarthria.4–10 Dysarthria is a motor speech disorder that may be associated with alterations in the regulation of phonation. The phonatory symptoms in MS are often masked by the pronounced scanning speech and abnormal prosodic patterns while speaking. However, the voice is a highly sensitive neurophysiological function, and a detailed study of vocal measures may reflect the status of nervous centers and pathways.7 Therefore, the study of the voice in carriers of neurological disorders aims at identifying relationships among pathophysiological mechanisms, lesion topography, and distinct phonatory alterations, for example, dysphonic symptoms. In MS, this type of finding may contribute toward maximizing the functional abilities of affected individuals. The objective of the present study was (1) to assess the presence of dysphonic symptoms in multiple sclerosis patients and (2) to compare quantitative acoustic parameters (fundamental frequency, jitter, shimmer, standard deviation of fundamental frequency, glottal noise level, frequency, and amplitude of tremor) in multiple sclerosis patients and normal individuals.
Only MS patients in the relapsing-remitting category as defined by Poser et al11 were included in the study. Exclusion criteria included the presence of vocal disorder prior to the appearance of neurological or clinical symptoms, previous larynx microsurgery, recent episode of endotracheal intubation, primary or metastatic tumor of the larynx, lung, or mediastinum, diagnosis of respiratory disease (acute/chronic), and any associated neurological diagnosis. The difference in gender distribution between the groups was statistically controlled. The study protocol was approved by the Ethics Committee at HCPA. All patients signed a consent form after receiving information about the objectives of the study and about the procedures to be carried out. During the data collection period, monthly meetings were held with the medical team in charge of the MS patients to review diagnostic criteria and to clarify possible doubts regarding inclusion of new patients. The acoustic evaluation of each patient was individually reviewed after the sample was recorded. Control of environmental noise was carried out following standard procedures.12 Data were entered into the database twice; a second analysis was carried out for 20% of the subjects, selected at random.
METHODS
Measurement of study variables A vocal sample was collected using a microphone (Optimus unidirectional microphone IMP.600, Philippines, 33-3017) connected to a computer (Acer, Extensa 368) through a preamplifier (Tiger Electronics, Seattle, WA). A computerized acoustic analysis of the samples was carried out using the software Dr. Speech Sciences, Version 3.2 (Tiger Electronics, Seattle, WA). This software was used to obtain quantitative parameters and graphic representations. For collection of vocal samples, individuals were asked to emit a sustained vowel /a / (as in father), trying to maintain the usual frequency and intensity. The microphone was placed on a microphone stand at a distance of 10 cm from the individual’s mouth. The rate used to record the signal was 44,100 Hz. In the present study, dysphonic symptoms were assessed using spectrographic analysis of wide and narrow bands. In the wideband analysis, the following items were scored: peak periodicity, darkness of
An 8-month controlled, cross-sectional study was designed to assess dysphonic symptoms in MS. The sample was composed of 30 individuals with multiple sclerosis, selected from among the patients in follow-up at the neurology service at Hospital de Clı´nicas de Porto Alegre (HCPA), a teaching hospital in southern Brazil. A control group was set up with 76 HCPA employees without neurological manifestations or psychiatric diagnoses, and without any chronic organic diseases. Socioeconomic status was similar for the MS and the control groups. Age ranged from 20 to 55 years. To calculate the size of the sample, the following criteria were observed: significance level (α) of 5%; power of 80% (β); and a control-case ratio of 2:1. This ratio was established because we expected dysphonic symptoms to be more frequent in the MS population in relation to controls. The exposure of normal individuals was estimated at 15%. Journal of Voice, Vol. 18, No. 3, 2004
MULTIPLE SCLEROSIS PATIENTS the contour, and presence of noise in high frequencies and in the formant region. In the narrow band, the presence of harmonics up to 4000 Hz as well as the regularity and definition of the contour were assessed. If three out of four items were altered in the wideband analysis, the examination was classified as pathologic; in the narrowband analysis, two out of three were altered. The spectrograms for the two bands were later compared. Dysphonic symptoms were diagnosed when both examinations were classified as pathologic.13 Assessment of the examinations was performed by a voice specialist. The examiner was blinded for the presence or absence of the study factor (MS). After that, fundamental frequency, jitter, shimmer, and noise measures were extracted from the samples. The software employed in the present study (Dr. Speech Sciences, Version 3.0)12 defines as normal jitter values up to 0.5% and shimmer values up to 3.0%. Glottal noise is considered to be present at an acoustic spectral level of ⫺10 dB. A maximum standard deviation of 3.0 Hz in relation to frequency is considered. The reference values for frequency were 80 to 150 Hz for males and 150 to 250 Hz for females.14 Statistical analysis The statistical analysis was carried out using the Statistical Package for the Social Sciences (SPSS, Chicago, IL).15 For dysphonic symptoms (qualitative variables), associations were assessed using Mantel-Haenszel’s χ2test, with Yates correction or the Fisher exact test when necessary. A p ⭐ 0.05 was considered to be significant. Jitter, shimmer, glottal noise, fundamental frequency, and frequency deviation (continuous variables) were analyzed using ANOVA or Student t test for independent samples or for dependent samples, when necessary. Multivariate analysis (MANOVA) was used to control confounding variables. Finally, a logistic regression model was used to analyze dependent variables both for MS and the presence of dysphonic symptoms. The model included the independent variables that presented differences in the MANOVA or χ2univariate analysis. RESULTS The spectrographic analysis of wide and narrow bands revealed that the frequency of dysphonic
343
symptoms among MS patients was 70% (n ⫽ 21/30) and, among controls, 33% (n ⫽ 25/76) (p ⫽ 0.01). Figure 1 shows the wideband and narrowband spectrogram of an MS patient with dysphonic symptoms in comparison with a normal individual. The demographic characteristics of the sample and the results of acoustic analysis appear in Table 1. Fundamental frequency and fundamental frequency deviation were higher in MS patients in comparison with the control group (p ⫽ 0.01). Because of the known importance of the gender variable for several aspects of the voice, the impact of this variable was analyzed separately. The controlled analysis of gender in individuals with and without multiple sclerosis is presented in Table 2. Jitter values were higher in MS men than in all other groups. Fundamental frequency deviation was higher in normal men than in normal women. In addition, this parameter was almost twice as high in MS men when compared with MS women, who presented values that were similar to those of normal women. The same was observed for glottal noise. To enable a more dynamic assessment of the simultaneous effects of the variables under study, a logistic regression model was set up having exposure to MS as the dependent variable, and gender, dysphonic symptoms, fundamental frequency, shimmer, noise, and fundamental frequency deviation as independent variables. After this analysis, dysphonic symptoms and fundamental frequency remained in the model (Table 3). From these, dysphonic symptoms presented a significant association with the dependent variable exposure to MS. Thus, the odds for having MS were 2.2 times higher in individuals presenting dysphonic symptoms.
DISCUSSION In the present study, we observed a statistically significant association between MS and dysphonic symptoms, with a frequency of 70% among subjects found to have MS. This finding is most probably explained by alterations in periaqueductal gray matter, which are common in MS.3,5 Such alterations have a major impact on phonation and are characterized by neuronal concentration and projections of abdominal, intercostal, pharyngeal, palatal, lingual, Journal of Voice, Vol. 18, No. 3, 2004
344
´ ET AL ADRIANA VE´LEZ FEIJO
FIGURE 1. Left. Narrowband analysis. Narrowband spectrogram of an MS patient with dysphonic symptoms (A) in comparison with a normal individual (B). Observe the lack and irregularity of harmonics in the MS patient. Right. Wideband analysis. Wideband spectrogram of an MS patient with dysphonic symptoms (C) in comparison with a normal individual (D). Observe the aperiodicity and a lack of darkness in the contour of the MS patient.
facial, masticatory, and laryngeal motoneurons into the nucleus retroambigualis.14,16,17 Even though the jitter values for both MS and normal men were within the normal range, our results suggest that the effect of jitter may be increased in MS men. The variability of muscle contraction resulting from the triggering of the motor unit in MS is one possible explanation for this. It is important to stress that chemical and electrical discharges, as well as the process of recovery of the myoneural junction, are faster than the recovery of the mechanical process.18 Therefore, the slowing down of the electrical impulse through the demyelination process Journal of Voice, Vol. 18, No. 3, 2004
in MS may cause an alteration of the recovery synchrony between the aspects mentioned previously, leading to a larger variation in the cycle-to-cycle frequency, ie, short-term variation or jitter. However, this finding must be interpreted with care, because jitter was not different when women with and without MS were compared. Also, it is important to stress that jitter, shimmer, and noise were not associated with MS in the logistic regression model, as also reported in other studies.7,19 Regarding long-term phonatory instability in MS subjects, both tremor amplitude and frequency were within the normal range for both genders. These
MULTIPLE SCLEROSIS PATIENTS TABLE 1. Demographic Information and Results of Vocal Analysis§ in Patients With and Without Multiple Sclerosis
Variables Age Gender Male Female Frequency Jitter Shimmer Fundamental frequency deviation Glottal noise Tremor frequency Tremor amplitude
Multiple sclerosis n ⫽ 30
Controls n ⫽ 76
P
45.10 ⫾ 7.24
42.16 ⫾ 7.64
0.07*
9.00 21.00 175.53 ⫾ 39.68 0.26 ⫾ 0.17 2.94 ⫾ 1.84 2.30 ⫾ 0.95
53.00 23.00 150.71 ⫾ 39.05 0.24 ⫾ 0.09 2.58 ⫾ 1.06 1.61 ⫾ 0.87
0.01** 0.01* 0.27* 0.21* 0.01*
⫺12.37 ⫾ 4.45 1.72 ⫾ 0.94
⫺12.78 ⫾ 3.56 2.96 ⫾ 3.87
0.62* 0.09*
1.44 ⫾ 0.69
1.47 ⫾ 1.03
0.90*
*
One-way variance analysis or MANOVA. χ test with Yates correction. § Sound card capture. ** 2
results differ from those of previous studies reporting tremor values starting at 11.2 dB and 7.5 Hz.20 However, the present results may have been affected by limitations in the software used for assessment, because in general these measurements did not match our clinical observations. For example, tremor
345
frequency and amplitude values were altered even in individuals with a normal voice and with a normal spectrographic evaluation. In the present study, all MS patients belonged to the relapsing-remitting category11,21; therefore, the assessment of voice was not related to degree of disability. However, it is possible to observe that the mean time since diagnosis was similar for men and women with MS. Despite the fact that MS patients were not individually assessed in terms of degree of disability, the highest possible number of individuals was enrolled during the 8 consecutive months of this study; the subjects enrolled presented a comparable disease duration, which characterized a similar profile. In the spectrographic analysis, certain contour characteristics were observed, such as irregular harmonics, presence of noise, and reduced darkening of the contour. Previous studies report that dysphonia is a symptom rather than a disease per se.6,7,22 In the case of neurological disorders, dysphonic symptoms are often among the earliest symptoms developed. We believe that the strong association between dysphonic symptoms and MS described in the present study may help to demonstrate that health professionals must be alert to the need for patients to be referred for neurological examination, as well as changing the cultural and social values that define a normal voice. In addition, our results underscore the
TABLE 2. Analysis of the Interaction Between Variables According to Gender and Presence or Absence of Multiple Sclerosis* Multiple sclerosis (n ⫽ 30)
Controls (n ⫽ 76)
Variable
Males (n ⫽ 9)
Females (n ⫽ 21)
Males (n ⫽ 53)
Females (n ⫽ 23)
F
P
Age Disease duration Fundamental frequency Jitter Shimmer Fundamental frequency deviation Glottal noise Tremor frequency Tremor amplitude
47.11 ⫾ 3.52 6.63 ⫾ 4.87 142.43 ⫾ 27.36
44.24 ⫾ 8.28 7.40 ⫾ 6.69 189.71 ⫾ 35.78
41.40 ⫾ 7.83 – 133.58 ⫾ 27.34
43.91 ⫾ 7.02 – 191.96 ⫾ 31.64
29.5399
0.00
0.33 ⫾ 0.30 2.02 ⫾ 1.08 4.11 ⫾ 2.79
0.23 ⫾ 0.05 2.43 ⫾ 0.89 2.44 ⫾ 0.96
0.23 ⫾ 0.08 1.36 ⫾ 0.65 2.69 ⫾ 1.17
0.25 ⫾ 0.10 2.20 ⫾ 1.04 2.33 ⫾ 0.67
4.6453 2.2033 10.6975
0.00 0.09 0.00
⫺9.89 ⫾ 4.51 2.02 ⫾ 1.07 1.51 ⫾ 0.87
⫺13.44 ⫾ 0.08 1.59 ⫾ 0.88 1.41 ⫾ 0.62
⫺11.92 ⫾ 3.72 2.58 ⫾ 3.25 1.54 ⫾ 1.20
⫺14.84 ⫾ 1.99 3.87 ⫾ 5.03 1.28 ⫾ 0.38
5.5792 1.8272 0.4161
0.00 0.14 0.74
Significance level ⬍0.05. Repeated letters in the same parameter indicate a significant difference between the groups.
*
Journal of Voice, Vol. 18, No. 3, 2004
´ ET AL ADRIANA VE´LEZ FEIJO
346
TABLE 3. Results of the Logistic Regression Analysis for Exposure to Multiple Sclerosis* b
Standard error
⫺0.7861 ⫺0.0164 4.0501
0.3369 0.0085 1.5329
Variables not in model
Score
p
Gender Shimmer Noise Fundamental frequency deviation
0.9307 0.5559 0.9959 0.0609
Variables in model Dysphonic symptoms Frequency Constant
0.33 0.45 0.31 0.80
p 0.01 0.05 0.00
R
OR
CI 95%
⫺0.1656 ⫺0.1176
2.2 1.0
1.13–4.25 0.97–1.00
R 0.0000 0.0000 0.0000 0.0000
Partial correlation is expressed as R and ranges from ⫺1 to ⫹1. OR ⫽ odds ratio. CI 95% ⫽ 95% confidence interval.
*
consequences that disregarding small vocal alterations may have for the health of affected individuals. This finding may also contribute toward speeding the process of diagnosis of neurological diseases.
CONCLUSIONS Our results reveal an association between MS and dysphonic symptoms. MS tends to worsen performance of the acoustic parameters fundamental frequency deviation, noise, and jitter. Men with MS present with worse vocal performance than do women with MS. The phoniatric therapeutic process in MS patients should be reevaluated in terms of focusing on the phonation and resonance levels of the phonatory system, and in terms of adding vocal therapy methods, because of the large association observed between dysphonic symptoms and MS. This would enable more favorable communication conditions in these patients, and it would mitigate the limitations associated with MS.
REFERENCES 1. Charcot J. Lectures on the diseases of the nervous system. London: The New Sudenhem Society; 1877:1. 2. Bauer HJ, Hanefeld FA. Multiple sclerosis—Its impact from childhood to old age. London: Saunders; 1993:132–170. 3. Matthews WB, Compston A, Allen IV, Martyn CN. McAlpine’s multiple sclerosis. 2nd ed. London: Churchill; 1991: 43–105. Journal of Voice, Vol. 18, No. 3, 2004
4. Darley F, Brown J, Goldstein N. Dysarthria in multiple sclerosis. J Speech Hear Res. 1972;15:229–245. 5. Guberman A. An introduction to clinical neurology: Pathophysiology, diagnosis, and treatment. Boston: Little, Brown; 1994. 6. Hartelius L, Svensson P. Speech and swallowing symptoms associated with Parkinson’s disease and multiple sclerosis: a survey. Folia Phoniatr Logop. 1994;46:9–17. 7. Ramig AL, Titze I, Scherer R, Ringel S. Acoustic analysis of voice of patients with neurologic disease: rationale and preliminary data. Ann Otol Rhinol Laryngol. 1988;97: 164–171. 8. Poser CM. The epidemiology of multiple sclerosis: a general overview. Ann Neurol. 1994;36:180–193. 9. Beukelman DR, Kraft GH, Freal J. Expressive communication disorders in persons with multiple sclerosis: a survey. Arch Phys Med Rehabil. 1985;10:675–677. 10. Lethlean J, Murdoch B. Naming in multiple sclerosis: effects of disease course, duration, age, and education level. J Med Speech Lang Pathol. 1994;2(1):43–56. 11. Poser CM, Paty DM, Scheinberg L, et al. New diagnostic criteria for multiple sclerosis: guidelines for research protocols. Ann Neurol. 1983;13:227–231. 12. Huang D, Minifie FD. Dr. Speech science for Windows software and real speech—user’s guide. Seattle, WA: Tiger Electronics US Inc.; 1995. 13. Rontal E, Rontal M, Rolnick MI. Objective evaluation of vocal pathology using voice spectrography. In: Baken RJ, Daniloff RG, eds. Readings in clinical spectrography of speech. California: Singular; 1991554–563. 14. Norusis MJ. Statistical package for the social sciences SPSS (SPSS/PC) for IBM PC/XT/AT. Chicago, IL: SPSS Inc.; 1986. 15. Hammarberg B, Gauffin J. Perceptual and acoustic characteristics of quality differences in pathological voices as related to physiological aspects. In: Fujimura O, Hirano M, eds. Vocal fold physiology. California: Singular; 1995: 238–303.
MULTIPLE SCLEROSIS PATIENTS 16. Holstege G, Ehling T. Two motor systems involved in the production of speech. In: Davis PJ, Fletcher NH, eds. Vocal fold physiology—controlling complexity and chaos. California: Singular; 1996:153–169. 17. Zhang SP, Davis PJ, Bandler R. Brain stem integration of vocalization: role of the nucleus retroambigualis. J Neurophysiol. 1995;74:2500–2512. 18. Ludlow C, Bassich C, Connor N, Coulter D. Phonatory characteristics of vocal fold tremor. J Phonet. 1986;14: 509–515.
347
19. Titze IR. A model for neurologic sources of aperiodicity in vocal fold vibration. J Speech Hear Res. 1991;34:460–472. 20. Davis P, Zhang SP, Bandler R. Midbrain and medullary regulation of vocalization. In: Davis PJ, Fletcher NH, eds. Vocal fold physiology—controlling complexity and chaos. California: Singular; 1996:121–134. 21. Hanson D. Neuromuscular disorders of the larynx. Otolaryngol Clin North Am. 1991;24:1035–1051. 22. Zwirner P, Murry T, Woodson G. Phonatory function of neurologically impaired patients. J Commun Disord. 1991; 24:287–300.
Journal of Voice, Vol. 18, No. 3, 2004